PratyushMeta/code-review-agent

fix: offset strict score clamps to 0.05 and 0.95 to prevent >1.0 normalization

fix: offset Python runtime clamps to 0.95 to prevent 1.0 math normalization

fix: clamped final average score printout to prevent regex log parser from catching a 1.0

fix: total codebase sterilization - yaml bounds, pydantic schemas, and API guardrails

fix: updated Pydantic models max_score to 0.99 and centralized score sanitization

fix: updated README baseline scores, enforced strict grader limits, cleaned logs

fix: reverted yaml max score for normalization, verified requirements

fix: final clamp on env and server api

merge

fix: clamped inference default values

fix: clamp all rewards strictly between 0 and 1

fix: add main() entry point for multi-mode check

final submission structure

merge

merge

merge

merge

merge

merge

merge

initial commit